highlights in mineralogical crystallography · 2015-12-10 · highlights in mineralogical...

Thomas Armbruster, Rosa Micaela Danisi (Eds.)

HIGHLIGHTS INMINERALOGICALCRYSTALLOGRAPHY

"Highlights in Mineralogical Crystallography" presents a collection of reviewarticles with the common topic: structural properties of minerals and syntheticanalogues. It is a valuable resource for mineralogists, materials scientists,crystallographers, and earth scientists. This book includes:

u An introduction to the RRUFF database for structural, spectroscopic, andchemical mineral identification.

u A systematic evaluation of structural complexity of minerals.u ab initio computer modelling of mineral surfaces.u Natural quasicrystals of meteoritic origin.u The potential role of terrestrial ringwoodite on the water content of the Earth's

mantle.u Structural characterization of nanocrystalline bio-related minerals by electron-

diffraction tomography.u The uniqueness of mayenite-type compounds as minerals and high-tech

ceramics.

u Presents top researchu Provides state-of-the-art methods and techniquesu Addresses mineralogists, crystallographers and materials scientists

Prof. Thomas Armbruster, University of Bern, Switzerland; Dr. RosaMicaela Danisi, University of Bern, Switzerland.

xii, 201 pages, 27 Fig.5110 Tables

Hardcover:RRP *€ [D] 99.95 / *US$ 140.00 /*GBP 74.99ISBN 978-3-11-041704-3

eBook:RRP *€ [D] 99.95 / *US$ 140.00 /*GBP 74.99PDF ISBN 978-3-11-041710-4EPUB ISBN 978-3-11-041721-0

Print/eBook:RRP *€ [D] 149.95 / *US$ 210.00 /*GBP 113.98ISBN 978-3-11-041711-1

Date of Publication: November 2015

Language of Publication: English

Subjects:Condensed Matter Physics Crystallography Geology and Mineralogy Materials Sciences, other

*Prices in US$ apply to orders placed in the Americas only. Prices in GBP apply to orders placed in Great Britain only. Prices in € representthe retail prices valid in Germany (unless otherwise indicated). Prices are subject to change without notice. Prices do not include postageand handling. RRP: Recommended Retail Price. Order now! [email protected]

degruyter.com

mailto:[email protected]

Barbara Lafuente, R. T. Downs, H. Yang, and N. Stone1 The power of databases: The RRUFF project

1.1 Introduction

In today’s world, many people rely on the availability of databases to perform dailyactivities such as checking meteorological forecast, finding a recipe, or verifying thespelling of a word. These are examples of actions that anyone can do quickly, rou-tinely, and often at little cost, because of access to the Internet and the developmentof extensive databases.

The science community has always relied on databases, which in the earlieryears were only available through collations of journals, books and data records. Theanalogic nature of these data resources dramatically limits the process of search-ing records and establishing relations between datasets. Nowadays, digital scientificdatabases and the World Wide Web have changed the way we can do science. Thisnew generation of databases not only provides access to information, but allows datasharing betweenmultidisciplinary research projects andmake it possible to cross-linkdiverse types of data.

Minerals are very important to both the world economy and research science,they supply the rawmaterials necessary for the development of modern societies andtheir study provides knowledge of our planet, the solar system, and the very natureof the universe. As such, their identification and characterization is critical to manyfields. However, only a few experts can accurately identifymineralswithout analyticalinstruments and their complementary databases. In general, minerals are uniquelycharacterized by both chemistry and crystal structure, or their derivative properties.For instance, the derivative property, color, can be a function of chemistry, or crystalhabit a function of structure.

The characterization and differentiation of minerals in the 19th century waslargely based on chemical composition and crystal forms. This led to the appear-ances of compendia of minerals and their diagnostic properties in the form of booksthat become the earliest publicly available databases of minerals, such as Dana’sSystem of Mineralogy (1837–1997) [1], Hintze’s Handbuch der Mineralogie (1897–1933)[2], or Goldschmidt’s Atlas der Krystallformen (1913–1923) [3] containing drawings ofall recognized crystals forms of minerals.

The discovery of X-ray diffraction techniques in 1912 brought about a dramaticchange in the way minerals could be characterized because the diffraction pattern ofa given mineral species remains consistent without regard to the crystal form, habit,color, or other variability in properties. This has led to a large amount of definitivedata that is used to help identify and characterizeminerals. Subsequent technologicalinnovations have increased the speed and volume of data acquisition, and therefore,the need of databases to organize and make available all this information.

Bereitgestellt von | De Gruyter / TCSAngemeldet

Heruntergeladen am | 30.11.15 18:06

2 | Barbara Lafuente, R. T. Downs, H. Yang, and N. Stone

In addition to chemical and diffraction techniques, spectroscopy has proven to bea valuable derivative methodology used to identify and characterize minerals. In par-ticular, Raman spectroscopy has shown itself to be a competitive technique becauseRaman spectra ofminerals depend fundamentally on their chemistries and structures.Moreover, this technique is generally nondestructive, requires little sample prepara-tion, it is quick and holds the promise of low cost. Infrared spectroscopy is anotherpopular technique that is complementary to Raman and is capable of providing dis-tinctive spectra suitable for search/match routines. Both spectroscopy techniques areundergoing major instrumentation changes towards portable and handheld systems(e.g. [4–7]).

Mineral identification using Raman spectroscopy is normally performed bysearch/match routines that compare the acquired spectrum with reference spec-tra from a database. The purpose of the RRUFF project is to develop such a Ramandatabase bymeasuring the chemistry, X-ray diffraction patterns, Raman, and infraredspectra of the known minerals and to make these data readily and freely available tothe scientific community, industry, and the general public. RRUFF database currentlycontains about 7000 mineral samples representing 3500 mineral species. The powerof the RRUFF database is that its data is not collected from disparate publicationsand procedures but rather is collected with the same methodology, on the same set ofsamples and with the same instrument.

Although they are not currently included in RRUFF, there are other analytical andspectroscopy techniques that aid in the characterization of minerals [8]. For exam-ple, Laser-Induced Break-down Spectroscopy (LIBS) provides chemical compositionby measuring all the elements present in a mineral sample. Mössbauer spectroscopyis also extensively used in mineralogy to examine the various valence states and thetype of coordination polyhedron occupied by iron atoms.

1.2 A brief history of the RRUFF project

Until recently, Raman spectroscopy was only accessible to specialists because it re-quired complex instruments worth hundreds of thousands of dollars. In 2003, theHamilton Sundstrand Sensor Systems company (CA, USA) and the University of Ari-zona (USA) were funded by NASA to develop a miniaturized high performance Ra-man system for Mars surface studies [9, 10]. The success of this project demonstratedthat inexpensive small Raman instruments had the capability to make Raman spec-troscopy a popular stand-alone field tool for routine and unambiguous identificationof minerals and other materials by non-experts and thus, would eventually becomeavailable to the broad public.

Mineral identification using Raman spectroscopy needs reference spectra, so an-ticipating the increasing use of the technique, Pr. Downs at the University of Arizonabuilt a Raman spectral database for minerals by obtaining representative samples of



1 The power of databases: The RRUFF project | 3

minerals that are also characterized and identified by X-ray diffraction and chemicalcomposition. There are a couple of models that can be used to fund database devel-opment. One is to charge an annual license, such as the funding model used by Inter-national Centre for Diffraction Data (ICDD) (http://www.icdd.com/). However, if thecost of Raman instruments become so inexpensive that it is less than the cost of a li-cense, then the ICDDmodel of funding is not reasonable. The RRUFF project databaseis based on the model of paying upfront for database development and providing thedata for free.

Mr.Michael Scott, the founding president of Apple Computers, provided the fund-ing to begin the project and create a proof of concept. He named it the RRUFF Project[11]. With this funding, the project developed procedures for validation and verifica-tion and the purchase of required instruments, including X-ray diffractometers andRaman spectrometers.

A requirement fromMr. Scott was that all the data collected by the RRUFF projectwould feed an open-access Raman spectral database not only for geologists and min-eralogists, but for the general worldwide public. Therefore, a website for the RRUFFdatabase (http://rruff.info) was built to provide free access to all the collected data,allowing users to view and download X-ray diffraction patterns, Raman and infraredspectra, electron microprobe data, references, photographs, and acquisition data ofwell characterized mineral samples. The RRUFF database is constantly reviewed andupdated, following strict protocols to ensure the quality and durability of the data.It currently receives ~80000 hits per week. Its success, together with the interest ofMichael Scott in mineralogy and Raman spectroscopy, has led him to also fund a sim-ilar free-access database, calledWURM (http://www.wurm.info/), which contains theresults of ab initio Raman spectra calculations [12].

It was also important in the early days of the RRUFF project to obtain a list ofthe known mineral species. With the help of the International Mineralogical Asso-ciation (IMA) and its nomenclature commission, an outreach committee of the IMAwas formed to develop and maintain an interactive and searchable list of the IMA-approved minerals. This list, referred in this paper as the IMA-List, is found at: http://rruff.info/ima/ (Fig. 1.1) and provides our preferred interface to the data of RRUFFas well as the Handbook of Mineralogy [13] maintained by the Mineralogical Societyof America, the American Mineralogist Crystal Structure Database (AMCSD) for struc-tural data [14–16], a reference library of publications and in some cases their pdf’s,a Mineral Evolution database for mineral localities and their ages, and links to otherdatabases that search on mineral names. In addition, and in order to promote privateuse of the IMA-List of minerals, the site includes an option to create additional linksto any other website that will accept queries based on mineral names. For example,we have a link to our mineral museum database. The site also offers a downloadablespell check dictionary file containing the mineral names from the IMA-List which canbe added to a user’s computer dictionary.




Fig.1.1:TheIMA-Listho

mepagelayout(http

://rruff

.info/ima/).At

thetopleftistheinteractivelistofIMA-approved

minerals,which

issearchableby

mineralname(e.g.quartz

)orb

ychem

istry

throughtheavailableperio

dictable.At

thetoprig

htisthelistofprovidedlinks

tootherm

ineraldatabase,including

ourcustomized

linktotheUn

iversityofArizonaMineralMuseum(UAM

M).




1.3 The RRUFF database

As of this moment, 3527 of the 4985 known mineral species have been incorporatedinto the RRUFF sample collection. As data from a sample is collected, it is postedinto the database with password restricted access, referred to as non-public access.After a review process, if the data appears to be representative of the sample, thenthe password restriction is removed and the data becomes publicly accessible. As aconsequence,measurements from only 2128mineral species are currently publicly ac-cessible. When possible, data from at least two samples of each species, ideally fromdifferent localities, are included in the database. Having multiple samples providesa means to confirm data and capture the chemical variability frequently observed inminerals. For instance, the database currently contains 42 records on the importantolivine forsterite-fayalite series, with associated Mg-Fe chemical variations. In total,data from 3791 samples are publicly accessible through the RRUFF database.

1.3.1 Sample collection and characterization

Mineral samples for the database were initially obtained from the University of Ari-zona Mineral Museum (UAMM) and from the purchase of the Cureton reference col-lection of 6500 specimens by Michael Scott. Currently, new mineral specimens aremostly obtained from donations by collectors, dealers, and museums around theworld. In 2013 alone, the RRUFF project received over 10000 samples through dona-tions, including two major ones: Mr. Rock Currier (~8000) and Dr. Robert Lavinsky(~2000).

We insert the samples of interest into the database by giving themauniqueRRUFFID (e.g. R070563) andwe scan and post the original labels, if they exist. We also recordimportant acquisition data, such as source and locality, according to the label. Afterthe record has been created, the next step is to confirm the identity of the sample. Wehave found that more than 25% of all the samples that we have studied have beenpreviously incorrectly identified on their labels. This failure rate illustrates the diffi-culty in determining the identity of a mineral, even for experts. Until the identifica-tion is confirmed, users should carefully suspect the correctness of the mineral name.The Status line in the top portion of each sample record reports the degree of the cer-tainty in the identification and the techniques used. Even though our goal is to haveall samples characterized by means of X-ray diffraction and chemical analysis, dueto limited resources and the cost of chemical analysis, we try to identify them first byX-ray diffraction because this equipment is in our laboratory. However, somemineralscan be difficult to be unambiguously identified by X-ray diffraction. For example, dif-ferent mineral species belonging to the garnet or amphibole groups may have similarcell parameters and because they are isostructural, their diffraction patterns will besimilar. In these cases, chemical composition is required. When using RRUFF data, it




is always important to review the Status field in order to know the degree of certaintywith which the minerals has been identified.

Most specimens in the RRUFF project are also documented with a macro photo-graph of the entire sample, intended to show host material, and a micro photographto illustrate the quality of themineral specimen under examination. Examples of bothmacro andmicro photos are shown in Fig. 1.2.

Fig. 1.2: Photosmacro (top) and photosmicro (bottom) of some RRUFF samples. From left to right:sigloite (R050576), bobdownsite (R050109), and xanthoconite (R080013).

1.3.1.1 Chemical analysisThe chemical composition is mainly determined with electron-microprobe analysis(EPMA) complemented, when necessary, with inductively coupled plasma mass-spectrometry (ICP-MS) analysis for the measurement of light elements (lighter thanoxygen, such as B, Be, and Li).

Electron-microprobe analyses are conducted at the University of Arizona ElectronMicroprobe Laboratory at the Lunar and Planetary Laboratory (LPL) on a fully auto-mated CAMECASX50 or SX100Ultra instrument. A detailed description ofmicroprobeanalysis applied to geology andmineralogy can be found in Reed [17] and the calcula-tions to determine the empirical chemical formula are explained in detail in Deer et al.[18]. The determination of light-elements (H, Li, Be, B) is done by inductively coupledplasma mass-spectrometry (ICP-MS) on a Thermo Scientific X-SERIES 2 QuadrupoleInductively Coupled Plasma Mass Spectrometer (Q-ICPMS).

The chemical analysis for eachmineral is posted to the RRUFFwebsite along witha photo of the image of the polished fragment used in the probe. An Excel file con-




Fig. 1.3: Chemistry section of lefontite (R140428) showing the measured chemistry, photo of SEMimage (link to a larger high-resolution image), and the access to the Excel spreadsheet.

taining all the probe data is also provided and it can be downloaded andmanipulatedby anyone who is interested in the results or to use as a template for other analyses(Fig. 1.3).

The IMA-List provides an interactive interface to search forminerals basedon theirideal chemistry using a periodic table (Fig. 1.4). With this tool, the search can be donenot only by elemental composition but also by valence state (i.e. Fe3+) and strings (i.e.H2O, OH), which results in a single selected mineral, or a set of minerals.

Fig. 1.4: Periodic table interface in the IMA-List. The example illustrates a searching based on min-erals that contain the elements Mn, P, O and Ca, the strings Fe3+ and H2O, and exclude Mg. Thesystem retrieves seven minerals meeting these conditions: bederite, jahnsite–(CaMnFe), jahnsite–(CaMnMn), keckite, wicksite, wilhelmvierlingite, and zodacite.

1.3.1.2 X-ray diffractionAll the RRUFF samples are examined by either powder or single-crystal X-ray diffrac-tion. When there is sufficient sample then powder X-ray diffraction is the method ofchoice because it produces a pattern that matches the established gold standard for-mat for identifying minerals. The reader can refer to Lavina et al. [19] for an extended




description on the modern X-ray diffraction methods in mineralogy and geosciences.For powder X-ray diffraction analysis, the RRUFF project uses a Bruker D8 Advancewhere a powdered sample is examined from 5° to 90° 2-theta at 2.0 seconds per 0.010°step with Cu radiation.

The diffraction patterns are compared against the ICDD using the EVA search/matchmodule provided by Bruker. Eachmeasured diffraction pattern is then indexedand cell parameters are refined with CrystalSleuth [20] using h k l values from AM-CSD, as computed by the module XPOW [21] included in the XtalDraw software [22].The cell parameter refinement in CrystalSleuth is based upon determining the posi-tions of isolated diffraction peak Kα1 and Kα2 doublets and fitting the set of positionswith a non-linear least-squares algorithm [20, 22] (Fig. 1.5). The peaks are fit using aχ2 minimization routine known as the Levenberg-Marquardt Method similar to thatdescribed in Press et al. [23].

On the RRUFF website, the diffraction pattern is displayed and the profiles are postedas XY-ASCII files, along with the refined crystallographic parameters.

Fig. 1.5: CrystalSleuth interface showing an example of fit and index processes of an experimentalpowder X-ray diffraction pattern of an adamite sample.

If the identification or refinement by powder X-ray diffraction is ambiguous due to,for instance, extreme peak overlap, or the amount of sample is too small for powderwork, then single-crystal X-ray diffraction is the preferred technique. The analyses areconducted using a state-of-the-art Bruker X8 diffractometer with Apex2 CCD detectorand Mo radiation.

In this case, the mineral is identified on the basis of its cell parameters andsearch/match is performed using the IMA-List, which offers a routine for search-




Fig. 1.6: Cell parameter search control interface in the IMA-List.

ing cell parameter extracted from the literature and from the RRUFF and AMCSDdatabases. It also provides a tolerance of 1%, 10%, or any value chosen by the userto the cell parameters to enable a search within a range of values (Fig. 1.6).

With single-crystal diffraction data, a theoretical powder diffraction pattern iscomputed using XPOW (Fig. 1.7) and it is posted in the database along with its as-sociated XY-ASCII file.

Fig. 1.7: Theoretical powder X-ray diffraction pattern of iranite [24] generated by XPOW.




1.3.1.3 Raman spectroscopyRaman spectroscopy is based upon the inelastic scattering of light due to its interac-tion with the vibrational modes of the target sample [25, 26]. During the analysis, thetransfer of energy of the incident photons (gain or loss) results in a spectrum of energyshifts characteristic of the chemistry and structure of the compound, and therefore,may provide a fingerprint that can be used in the identification of most minerals.

Interestingly, Griffith [27] studied the Raman spectra of the major rock-formingminerals in order to understand the effects of SiO4 condensation, and claimed thatRaman is “unlikely to be useful for the identification of silicate minerals”. Neverthe-less, the RRUFF project has shown that 80%of the samples examined can be correctlyidentified by their Raman spectra.

In general, four Raman spectra are collected from unoriented samples with 532(green) and 785/780 (red) nm lasers in a Thermo-Almega micro Raman system. Theinstrument is equipped with an Olympus BX microscope with 10X, 50X, and 100Xobjective lenses. Whenever possible, the measurements are collected using 50X. Thelasers are partially polarized with 4 cm−1 of maximum resolution and a spot size< 1 μm.High-resolution narrow scans, usually from 70–1400 cm−1, and low-resolutionwide scans usually from 70–4000 or 70–6500 cm−1 are collected with the 532 and785/780nm lasers from exactly the same spot in the sample and until the signal tonoise ratio is optimal.

The high-resolution narrow scans are used as reference spectra in the search/match library installed with CrystalSleuth (Fig. 1.8), because the strongest main Ra-manpeaks representative of a sample usually fallwithin this region. TheCrystalSleuthsearch/match routine treats the background-removed pattern as a multi-dimensionalvector (of norm 10) where an interpolated intensity (y-axis) at every twowavenumbers(x-axis) is treated as the coordinate value. We intentionally view the pattern and eachof the spectra within our search-library as definedwithin the same vector-space. A dotproduct reveals the spectra that are most similar. Because of the normalization, a per-fect match returns a value of 100. This process is repeated for the top matches; then,for greater resolution, the intensities are interpolated at every integral wavenumber.An ordered list of matching results with a confidence value (0 to 100) is listed leftof each result. The confidence is the calculated result of the dot products discussedabove, which indicates the coincidence between each pattern in the search/match li-brary and the spectrum of interest. As this software produces the closest matches tosamples in the installed RRUFF library, it is important to remember that even with ahigh confidence, this does not necessary imply amatch, just the bestmatchwithin thedatabase. In some cases, it is possible that the right reference is not even in the library.

The low-resolution wide scans are performed with both 532 and 785 nm lasers inorder to record other Raman peaks, such as the O-H stretchingmodes, and other spec-tral artifacts. With our instrument, only the spectra collected with the 532 nm laser re-veal the presence of O-H bands that are often located in the 2700–3700 cm−1 region, as




Fig. 1.8: CrystalSleuth interface showing an experimental unknown Raman spectra (in black), whichmatches the spectrum of the RRUFF record for prehnite R050410 (in blue).

shown in Fig. 1.9. The sensitivity of the silicon detector in the 785 nm laser is quite poorat higher wavenumbers which makes it useless for the detection of the O-H modes.

CrystalSleuth is used to correct the spectral baseline, using a subroutine from theRazor library (http://www.spectrumsquare.com) by Spectrum Square Associates, toremove cosmic rays events from patterns [28], to trim edges, to reverse X-axis display,and to visualize and compare multiple spectra. Both raw and processed data are in-cluded on RRUFF as XY-ASCII files.

The intensities of certain Raman peaks of a mineral may vary appreciably as afunction of sample orientation, depending upon its symmetry and the origin of theRaman modes (Fig. 1.10).

Such variations in intensity can significantly affect the success of the search/match routines. Thus, in certain cases, especially for rock-forming minerals, Ramanspectra collected in additional different orientations with respect to the polarizationdirection of the incident laser are collected. The analyses are conducted with an IRISRaman instrument built by Alex Goncharov and Victor Struzhkin of the GeophysicalLaboratory at the Carnegie Institution ofWashington. It uses a tunable 100mWAr-ionlaser (usually at 514.532 nm) and a Jobin Yvon Spex HR 460 spectrometer equippedwith a liquid nitrogen cooled Princeton Instruments 1152 × 256 pixel CCD detector.We use a 1200 groves mm−1 grating centered at 530.4 nm and collect data with RoperInstruments Winspec/32 software. Samples can be oriented onto a goniometer stageand rotated precisely into position, or spun at speeds up to 720 degrees per sec.

Some mineral records also provide a pdf file with the description of the Ramanactivemodes corresponding to themineral species (e.g. beryl, R040002). The analysis




tens

ityAr

bitr

ary

InA

100 400 700 1000 1300 1600 1900 2200 2500 2800 3100 3400 3700 4000Raman Shi� (cm-1)

Fig. 1.9: Low-resolution Raman spectrum of tetrawickmanite, Mn2+Sn4+(OH)6 (R100003), collectedwith the 532 nm laser, illustrating the O-H stretching vibrations Raman bands between 2700 and3700 cm−1.

is done by using the crystal structure data from AMCSD and the SAM routine at theBilbao Crystallographic server (http://www.cryst.ehu.es) [29].

1.3.1.4 Infrared spectroscopySimilarly to Raman spectroscopy, the infrared spectrum can provide additional infor-mation and a quickway to identify amineral, assuming that a database is available forcomparison. The two techniques are complementary because while Raman scatteringrequires a change in polarizability with vibration, infrared requires a change in dipolemoment. Therefore, in molecules with different elements of symmetry, certain bandsmay be infrared active, Raman active, both or neither. When sufficient powder sam-ple exists, the infrared absorption spectra are collected at the California Institute ofTechnology (Caltech) in Dr. George Rossman’s lab. The analyses are performed usinga SensIR Durascope on a ThermoNicolet Magna 860 FTIR (wavelength 4000–375 nm).The collected data is included as XY-ASCII file. The RRUFF project does not include asearch/match routine for identification of minerals by infrared spectroscopy. Addi-tional details can be found at George Rossman’s website at Caltech. The user can ac-cess the Mineral Spectrosocopy Server (http://minerals.gps.caltech.edu/) where visi-ble, near-infrared, and infrared absorption spectra are available for a number of min-




tens

ityAr

bitr

ary

InA

100 200 300 400 500 600 700 800 900 1000 1100 1200Raman Shi� (cm-1)

Fig. 1.10: Raman spectra of topaz (R040121) in two different orientations. The lower spectrum (or-ange) was recorded with the laser parallel to the b-axis, polarized parallel to c, while in the upperspectrum (blue) the laser is parallel to the c-axis, polarized parallel to b.

erals, including rock-forming minerals, gem minerals, and other minerals of particu-lar interest.

1.3.1.5 Data from external sourcesSome rare minerals are very difficult to obtain by the RRUFF project. However, if theRaman and infrared spectra are available in the literature, we contact the authorsand invite them to share their data through the RRUFF database. Microprobe anal-ysis and crystal structure information, if available from the manuscript, are alsoincluded in the record. We thank all the authors for sharing their data with theRRUFF project. The list of minerals needed to complete the project can be foundat http://rruff.info/about/minerals_needed.php. If you have Raman or infrared spec-tra of minerals that are not currently represented in the database, we encourage youto contact us.

1.3.2 Reference library

The RRUFF project includes an extensive reference library of publications directlylinked to their associated minerals. For the most part, these articles are focused onspectroscopy, structure, and chemistry of minerals. The complete list of collaboratorscan be found at: http://rruff.info/about/about_publishers.php. Only when the publi-




cation has permission for free viewing, its pdf can be freely downloaded. In addition,with funding from the National Science Foundation (The AmericanMineralogist Crys-tal Structure Database, 2006–2009, National Science Foundation EAR-0622371), theRRUFF project scanned and made public the older pre-digital journals of a number ofsocieties including USA, Canada, Great Britain, France, Italy, Russia, and Japan. Forthose minerals with reported structure, the reference list also provides a link to theirAMCSD record. The access to the references is facilitated either through the samplerecord itself, which provides a list of related publications, or through the webpageinterface for Search References (http://rruff.info/rruff_1.0/reference_search.php ).

1.3.3 Search and access to mineral records

Each record in the RRUFF database can be accessed by different pathways. Throughthe database homepage, an interactive search procedure provides a series of text fieldsfor searching by mineral name, chemistry, or keywords. It also includes a periodictable interface to search by chemical composition (lookup button), which is done byselecting elements included in the chemistry (click element once), or elements thatneeds to be excluded from the chemistry (click element twice). For example, all silicateminerals can be obtained by selecting only Si and O. To obtain the SiO2 polymorphs,the rest of elements in the periodic table needs to be excluded. The Exclude all non-selected button excludes all of them automatically. The search hits can be sorted anddisplayed in different formats that can be set as default.

The mineral records can be also accessed by typing the name of the mineral orthe RRUFF ID in the website URL in the format: http://rruff.info/[mineral name orRRUFF ID]. For example, rruff.info/quartz returns a page listing all the public quartzsamples that have been analyzed by the RRUFF project, while rruff.info/R040031 dis-plays all the data collected for one of the specific samples of quartz. This type of ac-cess also facilitates a way for other websites or software to independently access datafrom the RRUFF webpages. The IMA-List uses the URL method to provide a link to themineral records in RRUFF. This link will be only active (black font) when the RRUFFdatabase contains at least a record of the selected mineral.

1.3.4 Software design and server infrastructure

The RRUFF database is built using the open-source database-management softwareMySQL, which runs on all platforms. MySQL is combined with the PHP scripting lan-guage to create the database structure, to populate andmaintain the database, and tosupport data retrieval. PHP not only securely supports the MySQL database but alsoit is the link between the database and the Web page.




The Web search interface is dynamically created with PHP using the Select com-mands to access the criterion tables in the MySQL database. We use an Apache Webserver running on a Linux platform, which is well suited for PHP and MySQL. Thesearchable fields in the user interface include interactive HTML elements such as textboxes, drop-down list boxes, radio buttons, push buttons, etc., which are createdwithPHP. The results of the queries populate the search criteria fields in the HTML form.

The RRUFF project spends considerable time and resources to ensure the avail-ability of the RRUFF database at all times. As the database is used worldwide, it is apriority for the project to provide users with a reliable tool that can be used at any timeandwhose data last over time. Server infrastructure used for the RRUFF database usesRAID (Redundant Array of Inexpensive Disks) for local storage to ensure data integritywithin the local archive. This ensures that a single or double disk failure in the systemwill not result in loss of data. This hardware is regularly maintained and upgraded toensure performance and availability.

1.4 Experimental observations during collection of Raman spectra

An outcome of building the RRUFF database is the opportunity to examine hundredsof minerals by Raman spectroscopy and to observe diverse factors that may affect thecollection of the data [30]. We summarize some of these factors in this section.

1.4.1 Thermal effects of laser power

During the collection of a Raman spectrum, the incident laser beammay heat the illu-minatedportion of the sample,which, in some cases, can cause phase transitions, lossof hydrationwater, or partial or total decomposition of the sample. The increase of thetemperature depends upon experimental conditions, such as excitation wavelength,laser power, and exposure time, as well as mineral properties, including size, chem-istry, and color. When heating effects are suspected, we typically lower laser power,modify exposure time, or defocus the laser beam. The downside of reducing the laserpower to its minimum value is that it also implies lowering the intensity of the Ramansignal and worsening the signal to noise ratio, therefore providing a lesser quality Ra-man spectra.

Usually small or thin isolated crystals tend to burn more easily because they arenot able to dissipate much of the heat. When we collect Raman data from the samesingle crystal used for X-ray diffraction analysis, the power of the laser is usually re-duced to its lowest value. Minerals with weak bonds, such as many of the sulfides orhighly hydrated minerals, also tend to burn more easily even under the lowest laserintensity setting. Cinnabar (HgS) provides an examplewhere the high intensity 514nmlaser produces the spectrum of sulphur (S). Opaque phases, such as iron oxides, ab-




sorb more of the incident radiation making them also more susceptible to burning.Interestingly, even though the 532 nm laser is higher energy than the 785 nm laser, wehave found caseswhere amineral burns under 785 nmbut not under 532 nm (i.e. para-montroseite, VO2). Visual evidence such as dark brown to black craters of the size ofthe beam spot or discoloration of themineral surface are usually indicative of burningand dehydration, respectively.

Another not so often effect related to the heating of the sample by the incidentlaser is that while the spectra remains essentially the same, the peak positions shiftto lower wavenumbers with increasing the energy of the laser. We detect this phe-nomenon when we compare the spectra from the 532 and 785 nm lasers and correctthe lower wavenumber spectrum by lowering the intensity of the laser.

1.4.2 Effects of the incident laser wavelength

While the position of the peaks in theRaman spectrum remainsmostly constant underdifferent excitationwavelengths, their intensities can vary. Aswehave seen,we collectRaman scans using two different laser excitations from the same spot on the sampleand in the same orientation. Therefore, the high-resolution narrow scans provide a

tens

ityAr

bitr

ary

InA

100 200 300 400 500 600 700 800 900 1000 1100 1200Raman (cm-1)

Fig. 1.11: High-resolution Raman spectra of the zeolite edingtonite (R040003), collected with both532 (green line) and 785 (red line) nm lasers, illustrating an example of spectra whose patterns areessentially identical for both incident lasers. The sample was in identical position for both datacollections, which eliminates the variation in intensities caused by orientation effects.




tens

ityAr

bitr

ary

InA

100 200 300 400 500 600 700 800 900 1000 1100 1200Raman Shi� (cm-1)

Fig. 1.12: High-resolution Raman spectra of enstatite (R040093), collected with both 532 (green line)and 785 (red line) nm lasers, illustrating an example of spectra whose patterns differ for both lasers.The sample was in identical position for both data collections, which eliminates the variation inintensities caused by orientation effects.

measure of how the spectra change with incident wavelength. We have seen that forsome of the minerals, these two spectra may be the same or very alike, as in Fig. 1.11,but for others the intensities of identical modes can be different, as in Fig. 1.12.

1.4.3 Fluorescence

It has been found that about 16% of the samples examined in the RRUFF project pro-duced spectra that cannot be used for their identification, mostly due to the pres-ence of broadband fluorescence or luminescence peaks dominating the Raman spec-trum. Of these, half of them cannot be identified with either of the two lasers (532 and785 nm) (Fig. 1.13), while the other half can be identified only with one of the lasers(Fig. 1.14).

Usually, moving to longer wavelengths of excitation can significantly reduce flu-orescence [31] with the tradeoff of reducing the Raman signal because of the relationIRaman ~1/λ4 [25]. However, we have observed that some minerals fluoresce under 785nm but not under 532 nm, as shown in Fig. 1.14.




tens

ityAr

bitr

ary

InA

200 300 400 500 600 700 800 900 1000 1100 1200Raman Shi� (cm-1)

Fig. 1.13: High-resolution Raman spectra of loparite-Ce (R070251), collected with both 532 (greenline) and 785 (red line) nm lasers, illustrating an example of spectra where both incident wave-lengths are dominated by fluorescence.

tens

ityAr

bitr

ary

InA

100 200 300 400 500 600 700 800 900 1000 1100Raman Shi� (cm-1)

Fig. 1.14: High-resolution Raman spectra of dioptase (R040028), collected with both 532 (greenline) and 785 (red line) nm lasers, illustrating an example of spectra where one of the incident wave-lengths (780 nm laser) is associated with a poor pattern dominated by fluorescence.




tens

ityAr

bitr

ary

InA

100 600 1100 1600 2100 2600 3100 3600Raman Shi� (cm-1)

Fig. 1.15: Low-resolution Raman spectrum of zoisite (R060567), collected with the 785 nm laser,illustrating the significant luminescence features between 1200–1700 cm−1. Such features wereobserved in the spectra of all the Ca-containing minerals from the deposits in Umba Valley, Tanzania[30]. The study proposes that these features are due to Nd3+ luminescence centers.

Some minerals may exhibit obvious luminescence peaks, as in Fig. 1.15, often due tothe presence of transition metal or rare-earth ion centers [32], which in some situa-tions, it can be used as a signature for the locality of the sample [30].

If the luminescence features are very intense, as in Fig. 1.16, they can obscure theRaman peaks of the particular mineral.

1.4.4 Presence of inversion center

Of the 4967 known mineral species, twenty-seven of them cannot produce a Ramansignal because every atom is positioned on an inversion center. These include miner-als that crystallize with the rocksalt (16) or copper (11) structures. Inminerals contain-ing an inversion center, the change in polarization of the atoms due to the incidentbeam is negated because the distortion of the electron cloud in one direction is equalto the distortion of the cloud in another direction. As a result, the Raman spectrumhasintensity equal to zero. We have nevertheless collected the spectra for theses mineralsand included them in the database (Fig. 1.17) because the absence of Raman activityalso reveals important information about the sample. It is possible that some of thesetwenty-seven minerals may show weak Raman peaks. We interpret this as indicating




tens

ityAr

bitr

ary

InA

0 1000 2000 3000 4000 5000 6000Raman Shi� (cm-1)

Fig. 1.16: Low-resolution Raman spectrum of unoriented topaz (R060024), collected with the 532 nmlaser, illustrating intense luminescence features (4000–5200 cm−1) that mask the Raman peaksunder 1200 cm−1 (Fig. 1.10 shows the Raman bands of topaz in that specific region).

surface alteration. For instance, the oxidized surface of galena (PbS, rocksalt struc-ture) may show weak peaks corresponding to minium (Pb3O4).

1.4.5 Metamict minerals

Metamict minerals are materials that were initially crystalline, but that have lost theircrystallinity, partial or total, due to radioactive decay [33, 34]. Zircon [35], allanite[36], and titanite [37] are some examples of minerals that may exhibit metamictiza-tion. By annealing an unidentified metamict mineral at the appropriate temperature,its crystallinity can usually be restored and the sample identified by X-ray analysis.The RRUFF project studied eight metamict samples before and after heating to ex-plore the effects on their Raman spectra. For example, the metamict gadolinite-(Y)(R060856) was identified and its Raman obtained after heating the sample at 1000ºCduring 23 h [38].

Radiation damage produces change in the chemical and structural properties ofminerals, and therefore, it can have dramatic effects on the Raman spectrum of a sam-ple including loss of peak intensity, increase in peak width, and shifts in peak fre-quencies. For example, Nasdala et al. [39] calibrated the degree of metamictization innatural zircon crystals (ZrSiO4) by measuring the broadening and shift towards lower




tens

ityAr

bitr

ary

InA

200 300 400 500 600 700 800 900 1000 1100 1200Raman Shi� (cm-1)

Fig. 1.17: High-resolution Raman spectra of halite (R070534), collected with both 532 (green line)and 785 (red line) nm lasers, illustrating the absence of Raman activity in minerals with every atompositioned on an inversion center.

wavenumbers of the Raman band at ~ 1008 cm−1 (ν3SiO4). Fig. 1.18 shows the vari-ability in the position andwidth of the ν3SiO4 band in the Raman spectra of six zirconsamples collected for the RRUFF project.

1.5 Applications of the RRUFF project

The power of the RRUFF project lies in its capability to aid in mineral identificationand characterization. Records of well-characterized minerals can be used as a reli-able Raman reference, and the ancillary software, including CrystalSleuth, XtalDraw,and XPOW, provides visualization andmanipulation of data. Equally important to theRRUFFproject is that its database should alsoprovide the capability for inspiringotherresearch efforts. For instance, a search on March 2015 of the Science Direct website(http://www.sciencedirect.com) retrieved a listing of 339 manuscripts that use datafrom the RRUFF project. These publications come from a broad variety of disciplines,including mineralogy, crystallography, geology, archeology, planetary science, art,medicine,material science, physics, chemistry, environmental, biology, and forensics.

The abundant and diverse collection of Raman spectra, X-ray diffraction patterns,and chemical analysis permit the systematic examination and analysis of mineralgroups, chemical variations, phase transitions, etc. For example, using chemical anal-




tens

ityrb

itrar

y In

tAr

950 960 970 980 990 1000 1010 1020 1030 1040Raman Shi� (cm-1)

Fig. 1.18: Raman spectra in the 950–1040 cm−1 region of six unoriented zircons from the RRUFFdatabase collected with 532 nm laser, showing the shift and broadening of the ν3SiO4 mode:R050034 (blue), R100144 (purple), R050488 (green), R050286 (aquamarine), R100128 (red), andR050203 (orange).

yses of 40 various garnets from the RRUFF project, Henderson and Downs [40] showsthat the Raman spectra of silicate garnets are correlated to their empirical chemicalcompositions and can be used to estimate the compositions of unknown garnets towithin 3% atoms per formula unit, based on measurements of sixteen cations.

The collection of Raman spectra in the RRUFF database has had a profoundimpact on the gemology community. The inexpensive and non-destructive natureof Raman spectroscopy makes it a suitable technique for its use in this field, es-pecially given the perfection of gemstones. The Raman spectrum of gemstones isnot only used for their identification but also to determine their provenance (Fig.1.15), as well as natural or synthetic origin and possible treatments applied to them[41]. For example, Lowry et al. [42] acquire Raman spectra from 96 semi-preciousgemstones with the aim of classifying them with reference spectra from the RRUFFdatabase. The RRUFF database and CrystalSleuth are used by Kuehn [43] to study theapplication of Raman spectroscopy for the identification of gemstones. The work byJasinevicius [30] characterizes the vibrational and electronic features in the Ramanspectra of many gemstones in the RRUFF database. The Raman instruments devel-oped by M&A Gemological Instruments (MAGI) include the GemmoRaman software(http://www.gemmoraman.com/GemmoRaman-532.aspx) which provides a plugin




for the automatic export of collected spectra into CrystalSleuth for their identifica-tion.

The variability in the RRUFF Raman spectra collection, reflected in chemical com-position, laser wavelength, fluorescence, symmetry, site occupancy, and orientation,makes it suitable to use for building and testing mathematical models to automati-cally identify mineral phases from their Raman spectra. In Hermosilla Rodriguez etal. [44] they use one hundred randomly selected spectra from the RRUFF library totest an algorithm to identify mineral phases by Raman spectroscopy. In a recent pub-lication, Cochrane and Blacksberg [45] develop a method to identify individual Ra-man features inmineral mixtures by using 11572 Raman spectra downloaded from theRRUFF project. The development of these models is of special importance for plan-etary science as they will be necessary in the analysis of the data coming from theRaman instruments planned for future space missions (see for example [46, 47]). ARaman spectrometer has been chosen as a part of the science instrument payload ofthe European Space Agency 2018 ExoMars mission to target mineralogical and astro-biological investigations on the surface and subsurface of Mars [48, 49] and NASA hasjust announced that its 2020 Mars Rover payload [50] will carry two Raman spectrom-eters to determine fine-scale mineralogy and detect organic compounds.

The identification and characterization of mineral specimens for the RRUFFproject requires considerable research, mostly when something does not match withthe available data. The lack of a match may occur because it is a new mineral, be-cause its data was previously reported wrong, or because we do not have it yet in ourdatabase. So far, the IMA nomenclature commission has approved 21 new mineralsdiscovered and described by our group (e.g. [51–54]) with other 50 or so anomalousspecimens that require more careful examinations and currently remain non-publicin the database.

We also study crystal structures ofmineralswith previously unreported structures(e.g. [55–58]), or minerals with problems on their previously reported structural data(e.g. [59, 60]). In our experience, the greatest problem with previously reported cellparameters from the older literature is found where twinning or subcell/supercell re-lationswere not recognized, especially for data collectedwith point detectors. In somecases, the redefinition of the structure leads to a new ideal chemical formula for thatmineral. For instance, esperitewas first defined byMoore andRibbe [61] asmonoclinicwith a well-developed “superlattice”. However, Tait et al. [62] demonstrated that theso-called “superlattice” reflections were due to the presence of triple twins probablyresulting from ahigh-to-low temperature phase transformation, which led them to theredefinition of the ideal chemical formula from PbCa3(ZnSiO4)4 to PbCa2(ZnSiO4)3.In the case of the redefinition of the vladimirite, the work by Yang et al. [63] demon-strates that while the ideal chemical formula of this mineral was correctly reportedby Nefedov [64] as Ca4(AsO4)2(AsO3OH)⋅4H2O, it had been modified incorrectly asCa5H2(AsO4)4⋅5H2O by subsequent studies [65].




1.5.1 Crystal-chemistry analysis of Mars samples

The RRUFF project is playing an important role in the study of the chemical compo-sition of the mineral phases analyzed by the CheMin X-ray diffractometer [66] on theMars Science laboratory rover Curiosity in Gale crater,Mars. Using the refinedunit-cellparameters of samples analyzed at the sand dune, Rocknest [67, 68], Morrison et al.[69] estimates the empirical chemical formulas of the most abundant phases identi-fied, includingMg-Fe-olivine,Mg-Fe-Ca-pyroxene andNa-Ca-K-feldspar. For the study,a large dataset of unit-cell parameters in combination with their associated chemi-cal compositions were collected from the RRUFF project and the literature for each ofthese major phases. Moreover, by determining the chemical composition of the indi-vidual crystalline phases they were also able to estimate the bulk composition of thecrystalline materials as well as the amorphous component. This methodology will beapplied to each of the analysis performed by the CheMin instrument in the surface ofMars.

1.5.2 The Mineral Evolution database

The Mineral Evolution project [70] aims to understand how the episodes of planetaryaccretion and differentiation, plate tectonics, and origin of life lead to a selective evo-lution ofmineral species through changes in temperature, pressure, and composition.In order to investigate questions related to these mineralogical changes through timeand geological settings, the Mineral Evolution database [71] was created in partner-ship with the RRUFF project andMindat.org (http://www.mindat.org). The goal of thedatabase is to correlate the diversity of mineral species with their ages, localities, andother measurable properties. The RRUFF project serves as a platform to store and dis-play the data accessed through the IMA-List. Mindat.org supplies the database withmineral localities and their associated mineral species.

1.6 Future directions

The RRUFF database was designed to include measurements restricted to microprobeanalysis, Raman and infrared spectra, and X-ray diffraction patterns of minerals. Thisrigid structure prevents the addition of newmeasurements, such as Mössbauer, LIBS,or UV spectroscopy, without extensive and costly modification of the current softwareplatform. The custom programmed solution used by RRUFF also makes it impractica-ble for other researchers to use with their data.

These limitations have led to the development of the Open Data Repository [72],an open-access platform for data publication funded by the Mars Science LaboratoryCheMin team and NASA-Ames; developed by the University of Arizona’s Department




of Geosciences, and the Open Data Repository non-profit organization. The platformaims to provide researchers with a simple tool to publish any or all data related totheir research on material properties. The data publication system makes it possibleto share data publicly, to invite collaborators, and ensure research integrity.

With this system, users can create their own data repositories in their local serverfor their diverse data sets. Creating a data repository is a drag-and-drop procedurefacilitated through a web-based forms designer. No direct interaction with the under-lying database is required. The user can select and add different field types to theirdata sets and modify the design at any time to include new fields or file storage ar-eas. The user can then start creating data records that describe individual pieces ofdata. A plugin system allows data sets containing XY data to be graphed and plottedto HTML output. Plug-ins for rendering chemical formulas and for displaying imagesin galleries are also bundled with the system.

The platform has been in development for over two years and is undergoing betatesting by a number of research groups. NASA’s CheMin data from the Mars ScienceLaboratory is being published on the system as a proof of concept. Many new fea-tures and sharing tools will be developed under a new, 5-year contract with fundingfrom NASA’s SERA initiative and the Astrobiology Habitable Environments Database(AHED) [73] which will host data from a number of researchers in the astrobiologycommunity.

In the long term, the migration of the RRUFF database into this new platform willmake it possible to complement the current characterization of themineral specimenswith other techniques.

Acknowledgment: We are especially indebted to the support to the RRUFF Project byMr. Michael Scott, who provided four million dollars in cash and 1000 rare mineralsamples, as well as constant management advice and continuing encouragement.We also gratefully acknowledge support by the National Science Foundation EAR-0622371, Newmont Mining Corporation, Freeport-McMoRan Corporation, a StrategicUniversity Research Partnership from the Jet Propulsion Laboratory, and the ArizonaScience Foundation. This work is currently supported by the ChangshaMineral Show,NASA NNX11AP82A Mars Science Laboratory Investigations, and Science-EnablingResearch Activity NASA-Ames. We thank the hundreds of mineral enthusiasts, bothamateurs and professionals, who have contributed time and samples to the project.We also thank editors Rosa Micaela Danisi and Thomas Armbruster for their reviewsand constructive comments.




Bibliography

[1] Dana JD. A system of mineralogy. New Haven, CT, USA, Durrie & Peck and Herrick & Noyes,1837.

[2] Hintze C. Handbuch der Mineralogie. Leipzig, Germany, Verlag Von Veit & Comp, 1897–1933.[3] Goldschmidt V. Altas der Krystallformen. Heidelberg, Germany, Carl Winters Universitäts-

buchhandlung, 1913–1923.[4] Vandenabeele P, Castro K, Hargreaves M, Moens L, Madariaga JM, Edwards HGM. Compar-

ative study of mobile Raman instrumentation for art analysis. Anal Chim Acta 2007, 588,108–16.

[5] Jehlička J, Vítek P, Edwards HGM, Hargreaves MD, Čapoun T. Fast detection of sulphate miner-als (gypsum, anglesite, baryte) by a portable Raman spectrometer. J Raman Spectrosc 2009,40, 1082–1086.

[6] Vítek P, Ali EMA, Edwards HGM, Jehlička J, Cox R, Page K. Evaluation of portable Raman spec-trometer with 1064 nm excitation for geological and forensic applications. Spectrochim ActaA 2012, 86, 320–327.

[7] Kristova P, Hopkinson L, Rutt K, Hunter H, Cressey G. Quantitative analyses of powderedmulti-minerallic carbonate aggregates using a portable Raman spectrometer. Am Mineral2013, 98, 401–409.

[8] Henderson GS, Neuville DR, Downs RT. Spectroscopic methods in mineralogy and materialsciences. Washington, DC, USA, The Mineralogical Society of America, 2014.

[9] McIntosh B, Denton MB, Downs RT, Becker D. Smart Raman instrument for Mars Science Lab-oratory. Mars Exploration Program Assessment Group, Sep 10–11, Jet Propulsion Laboratory,Pasadena, CA, USA, 2003.

[10] McIntosh B, Denton MB, Downs RT. Process Raman on Earth and research Raman on Mars –How are they related? Pittcon, March 11, Chicago, IL, USA, 2004.

[11] Downs RT. The RRUFF Project: an integrated study of the chemistry, crystallography, Ramanand infrared spectroscopy of minerals. Program and Abstracts of the 19th General Meeting ofthe International Mineralogical Association in Kobe, Japan, 2006.

[12] Caracas R, Bobocioiu E. The WURM project – a freely available web-based repository of com-puted physical data for minerals. Am Mineral 2011, 96, 437–443.

[13] Anthony JW, Bideaux RA, Bladh KW, Nichols MC. Handbook of Mineralogy. Chantilly, VA, USA.Mineralogical Society of America, http://www.handbookofmineralogy.org/.

[14] Downs RT, Hall-Wallace M. The American Mineralogist Crystal Structure Database. Am Mineral2003, 88, 247–250.

[15] Clark CM, Downs RT. Using the American Mineralogist Crystal Structure Database in the class-room. J Geos Educ 2004, 51, 76–80.

[16] Rajan H, Uchida H, Bryan DL, Swaminathan R, Downs RT, Hall-Wallace M. Building the Amer-ican Mineralogist Crystal Structure Database: A recipe for construction of a small Internetdatabase. In: Sinha A K, ed., Geological Society of America Special Paper 2006, 397, 73–80.

[17] Reed SJB. Electron Microprobe Analysis and Scanning Electron Microscopy in Geology, 2nd

ed. New York, USA, Cambridge University Press, 2005.[18] Deer WA, Howie RA, Zussman, J. An introduction to the rock-forming minerals, 3rd ed. Lon-

don, UK, The Mineralogical Society, 2013.[19] Lavina B, Dera P, Downs RT. Modern X-ray diffraction methods in mineralogy and geosciences.

In: Henderson GS, Neuville DR, Downs RT eds. Washington, DC, USA, The Mineralogical Soci-ety of America, 2014, 1–31.

[20] Laetsch TA, Downs RT. Software for identification and refinement of cell parameters frompowder diffraction data of minerals using the RRUFF Project and American Mineralogist Crys-




tal Structure Databases. Program and Abstracts of the 19th General Meeting of the Interna-tional Mineralogical Association in Kobe, Japan, 2006.

[21] Downs RT, Bartelmehs KL, Gibbs GV, Boisen MB Jr. Interactive software for calculating anddisplaying X-ray or neutron powder diffractometer patterns of crystalline materials. Am Min-eral 1993, 78, 1104–1107.

[22] Bartelmehs KL, Gibbs GV, Boisen MB Jr, Downs RT. Interactive computer software used inteaching and research in mineralogy at Virginia Tech. GSA Fall Meeting, Boston, MA, USA,1993.

[23] Press WH, Teukolsky SA, Vetterling WT, Flannery BP. Numerical recipes in Fortran 77: the artof scientific computing, 2nd ed. New York, USA, Cambridge University Press, 1996.

[24] Yang H, Sano JL, Eichler C, Downs RT, Costin G. Iranite, CuPb10(CrO4)6(SiO4)2(OH)2, isomor-phous with hemihedrite. Acta Crystallogr 2007, C63, i122–i124.

[25] Rull F. The Raman effect and the vibrational dynamics of molecules and crystalline solids. In:Dubessy J, Caumon MC, Rull F, eds. London, UK, EMU 2012, 1–60.

[26] Neuville DR, de Ligny D, Henderson GS. Advances in Raman spectroscopy applied to Earthand material sciences. In: Henderson GS, Neuville DR, Downs RT, eds. Washington, DC, USA,The Mineralogical Society of America 2014, 509–541.

[27] Griffith WP. Raman studies on rock-forming minerals. Part I. Orthosilicates and cyclosilicates.J Chem Soc (A) 1969, 1372–1377.

[28] Hill W, Rogalla D. Spike-correction of weak signals from charge-coupled devices and its appli-cation to Raman spectroscopy. Anal Chem 1992, 64, 2575–2579.

[29] Kroumova E, Aroyo MI, Perez-Mato JM, Kirov A, Capillas C, Ivantchev S, Wondratschek H. Bil-bao Crystallographic Server: Useful databases and tools for phase-transition studies. PhaseTransit 2003, 76, 155–170.

[30] Jasinevicius R. Characterization of vibrational and electronic features in the Raman spectra ofgem minerals. Master Thesis, University of Arizona, Tucson, AZ, USA, 2009.

[31] Dubessy J, Caumon MC, Rull F, Sharma, S. Instrumentation in Raman spectroscopy: elemen-tary theory and practice. In: Dubessy J, Caumon MC, Rull F, eds. London, UK, EMU, 2012, 83–172.

[32] Gaft M, Reisfeld R, Panczer G. Luminescence spectroscopy of minerals and materials, Heidel-berg, Germany, Springer, 2005.

[33] Ewing RC. The crystal chemistry of complex niobium and tantalum oxides. IV. The metamictstate: discussion. Am Mineral 1975, 60, 728–733.

[34] Ewing RC, Haaker RF. The metamict state: implications for radiation damage in crystallinewaste forms. Nucl chem waste man 1980, 1, 51–57.

[35] Palenik CS, Nasdala L, Ewing RC. Radiation damage in zircon. Am Mineral 2003, 88, 770–781.[36] Janeczek J, Eby RK. Annealing of radiation damage in allanite and gadolinite. Phys Chem

Miner 1993, 19, 343–356.[37] Hawthorne FC, Groat LA, Raudsepp M, et al. Alpha-decay damage in titanite. Am Mineral 1991,

76, 370–396.[38] Lima de Faria J. Heat treatment of metamict euxenites, polymignites, yttrotantalites,

samarskites, pyrochlores, and allanites. Mineralogical Museum, section of the Junta de In-vestigações do Ultramar 1958, 3R, 937–942.

[39] Nasdala L, Wenzel M, Vavra G, Irmer G, Wenzel T, Kober B. Metamictisation of natural zircon:accumulation versus thermal annealing of radioactivity-induced damage. Contrib MineralPetrol 2001, 141, 125–144.

[40] Henderson RR, Downs RT. Determining chemical composition of the silicate garnets usingRaman spectroscopy. Master Thesis, University of Arizona, Tucson, AZ, USA, 2009.




[41] Bersani D, Lottici PP. Applications of Raman spectroscopy to gemology. Anal Bioanal Chem2010, 397, 2631–2646.

[42] Lowry S, Wieboldt D, Dalrymple D, Jasinevicius R, Downs RT. The use of a Raman spectraldatabase of minerals for the rapid verification of semiprecious gemstones. Spectroscopy2009, 24, 1–7.

[43] Kuehn JW. Raman and photoluminescence spectroscopy in gem identification. GIT proceed-ings Dec 8–9, Chiangmai, Thailand, 2014.

[44] Hermosilla Rodriguez I, Lopez-Reyes G, Llanos DR, Rull Perez F. Automatic Raman spectraprocessing for Exomars. In: Pardo-Igúzquiza E, Guardiola-Albert C, Heredia J, Moreno-MerinoL, Durán JJ, Vargas-Guzmán JA, eds., Berlin, Germany, Springer 2014, 127–130.

[45] Cochrane CJ, Blacksberg J. A fast classification scheme in Raman spectroscopy for the identi-fication of mineral mixtures using a large database with correlated predictors. IEEE Transac-tions on geoscience and remote sensing 2015, 53, 4259–4274

[46] Sobron P, Sobron F, Sanz A, Rull F. Raman signal processing software for automated identifi-cation of mineral phases and biosignatures on Mars. Appl Spectrosc 2008, 62, 364–370.

[47] Lopez-Reyes G, Sobron P, Lefebvre C, Rull F. Multivariate analysis of Raman spectra for theidentification of sulfates: Implications for ExoMars. Am Mineral 2014, 99, 1570–1579.

[48] Rull F, Maurice S, Diaz E, Tato C, Pacros A, RLS Team. The Raman Laser Spectrometer (RLS) onthe ExoMars 2018 Rover Mission. 42nd LPSC. Houston, TX, USA, 2011.

[49] Rull F, Sansano A, Díaz,E, et al. ExoMars Raman laser spectrometer for Exomars. Proc of SPIE2011, 8152, 1–13.

[50] NASA Announces Mars 2020 Rover Payload to Explore the Red Planet as Never Before. NASAPress Release 14–208, Jul. 31, 2014. (Accessed April 8, 2015, at http://www.nasa.gov/press/2014/july/nasa-announces-mars-2020-rover-payload-to-explore-the-red-planet-as-never-before/).

[51] Yang H, Downs RT, Evans SH, Morrison SM, Schumer BN. Lefontite, IMA 2014–2075. CNMNCNewsletter No. 23, February 2015, page 55; Mineral Mag 2015, 79, 51–58.

[52] Yang H, Downs RT, Evans SH, Pinch WW. Lavinskyite, K(LiCu)Cu6(Si4O11)2(OH)4, isotypic withplancheite, a new mineral from the Wessels mine, Kalahari Manganese Fields, South Africa.Am Mineral 2014, 99, 525–530.

[53] Tait KT, Barkley MC, Thompson RM, Origlieri MJ, Evans SH, Prewitt CT, Yang H. Bobdownsite,a new mineral species from Big Fish River, Yukon, Canada, and its structural relationship withwhitlockite-type compounds. Can Mineral 2011, 49, 1065–1078.

[54] Yang H, Jenkins RA, Thompson RM, Downs RT, Evans SH, Bloch EM. Markascherite,Cu3(MoO4)(OH)4, a new mineral species polymorphic with szenicsite, from Copper Creek,Pinal County, Arizona, U.S.A. Am Mineral 2012, 97, 197–202.

[55] Lafuente B, Downs RT, Yang H, Jenkins RA. Calcioferrite with composition(Ca3.94Sr0.06)Mg1.01(Fe2.93Al1.07)(PO4)6(OH)4•12H2O. Acta Cryst 2014, E70, i16–i17.

[56] Schumer BN, Downs RT, Domanik KJ, Andrade MB, Origlieri MJ. Pirquitasite, Ag2ZnSnS4. ActaCryst 2013, E69, i8–i9.

[57] Origlieri MJ, Downs RT. Schaurteite, Ca3Ge(SO4)2(OH)6•3H2O. Acta Cryst 2013, E69, i6.[58] Lafuente B, Yang H, Downs RT. Crystal structure of tetrawickmanite, Mn2+Sn4+(OH)6. Acta

Cryst 2015, E71, 234–237.[59] Andrade MB, Doell D, Downs RT, Yang H. Redetermination of katayamalite,

KLi3Ca7Ti2(SiO3)12(OH)2. Acta Cryst 2013, E69, i41.[60] Morrison SM, Downs RT, Yang H. Redetermination of kovdorskite, Mg2PO4(OH)•3H2O. Acta

Cryst 2012, E68, i12–i13.[61] Moore PB, Ribbe PH. A study of “calcium-larsenite” renamed esperite. Am Mineral 1965, 50,

1170–1178.




[62] Tait KT, Yang H, Downs RT, Li C, Pinch WW. The crystal structure of esperite, with a revisedchemical formula, PbCa2(ZnSiO4)3, isostructural with beryllonite. Am Mineral 2010, 95, 699–705.

[63] Yang H, Evans SH, Downs RT, Jenkins RA. The crystal structure of vladimirite, with a revisedchemical formula, Ca4(AsO4)2(AsO3OH)•4H2O. Can Mineral 2011, 49, 1055–1064.

[64] Nefedov EI. Report on new minerals discovered by him. Zap Vses Mineral Obshchest 1953,82, 311–317 (in Russian).

[65] Pierrot R. Contribution à la minéralogie des arséniates calciques et calcomagnésiens na-turels. B Soc Fr Mineral Cristallogr 1964, 87, 169–211 (in French).

[66] Blake D, Vaniman D, Achilles C, et al. Characterization and calibration of the CheMin miner-alogical instrument on Mars Science Laboratory. Space Sci Rev 2012, 170, 341–399.

[67] Blake DF, Morris RV, Kocurek G, et al. Curiosity at Gale Crater, Mars: Characterization andanalysis of the Rocknest sand shadow. Science 2013, 341, 1239505, 1–7.

[68] Bish DL, Blake DF, Vaniman DT, et al. X-ray diffraction results from Mars Science Laboratory:Mineralogy of Rocknest at Gale Crater. Science 2013, 341, 1238932, 1–5.

[69] Morrison SM, Downs RT, Blake DF, et al. Crystal-chemical analyses of Martian minerals inGale crater. 21st General Meeting of the IMA, Gauteng, South Africa, 2014.

[70] Hazen RM, Papineau D, Bleeker W, et al. Mineral evolution. Am Mineral 2008, 93, 1693–1720.[71] Hazen RM, Bekker A, Bish DL et al. Needs and opportunities in mineral evolution research.

Am Mineral 2011, 96, 953–963.[72] Stone N, Lafuente B, Downs RT, Blake D, Bristow T, Fonda M. The Open Data Repository’s Data

Publisher. AbSciCon, Chicago, IL, USA, 2015.[73] Lafuente B, Stone N, Downs RT, Blake D, Bristow T, Fonda M. The Astrobiology Habitable Envi-

ronments Database (AHED). AbSciCon, Chicago, IL, USA, 2015.



highlights in mineralogical crystallography · 2015-12-10 · highlights in mineralogical...

Documents